On Tuesday 04 March 2008 11:13, Stephen Graham wrote:
> Hi all,
>
> I would like to generate a structure-based multiple sequence alignment
> using 4 structures. I have already generated pairwise alignments for
> each 'pair' of structures (6 alignments in all). Is there a program
> out there that can take a number of aligned structures (or even just a
> number of pairwise sequence alignments) and calculate the 'best'
> multiple sequence alignment? Please note that there is absolutely no
> sequence conservation between these structures, making standard
> sequence-based alignment tools pretty useless.
My preference is for the CE-MC server. There are several sites that
offer this, or you can install it locally. Here's one:
bioinformatics.albany.edu/~cemc
Another one is Tcoffee:
http://tcoffee.vital-it.ch/cgi-bin/Tcoffee/tcoffee_cgi/index.cgi
When I've done head-to-head comparisons, I have found the output from
CE-MC to be preferable. In particular, in cases where you have two
structures A and B that have the same length but have an area in which
the structures differ, CE-MC will favor the alignment
AAAAAAaaa...AAAA
BBBBBB...bbbBBBB
whereas Tcoffee will report
AAAAAAaaaAAAA
BBBBBBbbbBBBB
For my purposes, the first is correct, even if aaa and bbb are
sequence-homologous.
--
Ethan A Merritt Courier Deliveries: 1959 NE Pacific
Dept of Biochemistry
Health Sciences Building
University of Washington - Seattle WA 98195-7742
